Acoustic detections per sex/month/area

analysis progress of the acoustic detections, master thesis Lotte Pohl, 27.2.2023

some info

  • Two tag serial numbers where only detected once: 1293296 and 1293301, and where thus omitted.

  • Females had a mean TL of 85.84 cm (\(\pm\) 10.26, n = 19), and males had a mean TL of 77.64 cm (\(\pm\) 3.75, n = 11).

  • Last time, we wanted to resolve further how detections might differ between sex and season.

  • I first divided the year into summer (from 1st of April to 30th of September) and winter (rest). During the winter period, however, there was not a single detection. So I then divided the detections by month.

  • Detections range from April to October.

  • Overall, 14 females were detected 10655 times and 4 males were detected 285 times.

  • I divided the study area into 3 parts: WS1 (the first array in the Westerscheldt), WS2 (the second, more east array in the Westerscheldt) and BPNS.

  • Within the BPNS, I grouped receivers more than 20 km away from the coast into offshore receivers and the ones closer than 20 km into coast receivers (analogous to Jolien Goossen’s Master Thesis).

maps

  • The circle size corresponds to the number of detections from each station each month. It is relative to the total number of detections.

  • The red circles are detections of females, the blue circles are detections of males.

I got salinity data from the ScheldeMonitor. They only reach until Vlissingen, unfortunately. So it roughly is the WS2 area. Below you see salinity averages. For some reason, there are no data for January, Febuary and March.

  • The circle size corresponds to the number of individuals detected at a station each month. It is relative to the total number of detected individuals.

  • The red circles are detections of females, the blue circles are detections of males.

barplots

The upper facet corresponds to females (f), the lower facet corresponds to males (m).

The upper facet corresponds to females (f), the lower facet corresponds to males (m).

abacus plot

All detections, the colour denotes the area and the symbol denotes the sex.

interpretations

  • We can see that females overall dominate the detections. By tagging 19 females and 11 males, we already integrated sex bias into the dataset. This bias further increased by only 4 males in contrast to 14 females being detected by a receiver.

  • One male was detected in the WS1 area. It must have swum into the array, because the sharks tagged in the WS (5km north of Cadzand) were all female. Between May and August, males are only detected in the BPNS. In the 4 months that males were detected, it was a different male each time (tag 311 in April, 315 in May, 316 in July and tag 295 in August).

  • In September, 2 females (tag 299 and 307) are detected in the WS2 area. There are only a couple of detections (38, on 22.09.2019 for tag 299 and 09.09.2019 for tag 307). Both these tags were also detected in the WS1 area in September 2019, tag 299 16 days before and 3 days after the detection in the WS2 area, and tag 307 1 day before and 7 days after. So tag 299 could have stayed in WS2 for 19 days, and tag 302 for 8 days.

residency index (RI)

  • The overall Residency Index (RI) is calculated by dividing the days of acoustic detections by the estimated battery time (for the ADST-V13TP it’s 518 days), for each tag.

  • The monthly Residency Index is calculated by dividing the days of detections in a month by the days of that month, for each tag.

  • I further subdivided the monthly RI by grouping the detections per tag and month into the three areas (WS1, WS2 and BPNS) to get a Residency Index per area and month.

The left facet corresponds to females (f), the right facet corresponds to males (m).

what I plan to do now

  • Home range analysis using horizontal Kernel Utilization Distribution (KUD).

  • what kind of statistics could I do on these data? I would like to see if there are significant differences between sex, area and month (could also be grouped into season, i.e. early/mid/late summer). I am not sure how to proceed, because we have different n for males and females, and overall a very small number of individuals. Do you have a statistical method in mind that could come in handy in this case?

  • In her master thesis, Jolien Goossens did a Correspondance Analysis (CA) on her monthly Residency Indices. What do you think of doing that here as well?

  • I have started the actel analysis and want to also use the RSP package which does home range analysis.

  • Overall, I would like to look at environmental variables (salinity, temperature, pH, …) and try to infer correlations with shark detections with a GLMM. But that would be 2nd priority after the home range analysis.